Available online at www.sciencedirect.com
R
Journal of Volcanology and Geothermal Research 126 (2003) 201^223
www.elsevier.com/locate/jvolgeores
Impact of the AD 79 explosive eruption on Pompeii, I.
Relations amongst the depositional mechanisms of the
pyroclastic products, the framework of the buildings
and the associated destructive events
Giuseppe Luongo, Annamaria Perrotta, Claudio Scarpati Dipartimento di Geo¢sica e Vulcanologia, Universita' Federico II di Napoli, Largo San Marcellino 10, 80138 Napoli, Italy
Received 26 February 2002; accepted 24 April 2003
Abstract
A quantitative and qualitative evaluation of the damage caused by the products of explosive eruptions to
buildings provides an excellent contribution to the understanding of the various eruptive processes during such
dramatic events. To this end, the impact of the products of the two main phases (pumice fallout and pyroclastic
density currents) of the Vesuvius AD 79 explosive eruption onto the Pompeii buildings has been evaluated. Based on
different sources of data, such as photographs and documents referring to the archaeological excavations of Pompeii,
the stratigraphy of the pyroclastic deposits, and in situ inspection of the damage suffered by the buildings, the present
study has enabled the reconstruction of the events that occurred inside the city when the eruption was in progress. In
particular, we present new data related to the C.J. Polibius’ house, a large building located inside Pompeii. From a
comparison of all of the above data sets, it has been possible to reconstruct, in considerable detail, the stratigraphy of
the pyroclastic deposits accumulated in the city, to understand the direction of collapse of the destroyed walls, and to
evaluate the stratigraphic level at which the walls collapsed. Finally, the distribution and style of the damage allow us
to discuss how the emplacement mechanisms of the pyroclastic currents are influenced by their interaction with the
urban centre. All the data suggest that both structure and shape of the town buildings affected the transport and
deposition of the erupted products. For instance, sloping roofs ‘drained’ a huge amount of fall pumice into the
‘impluvia’ (a rectangular basin in the centre of the hall with the function to collect the rain water coming from a hole
in the centre of the roof), thus producing anomalous deposit thicknesses. On the other hand, flat and low-sloping
roofs collapsed under the weight of the pyroclastic material produced during the first phase of the eruption (pumice
fall). In addition, it is evident that the walls that happened to be parallel to the direction of the pyroclastic density
currents produced during the second eruptive phase were minimally damaged in comparison to those walls oriented
perpendicular to the flow direction. We suggest that the lower depositional parts of the pyroclastic currents were
partially blocked (locally reflected) and slowed down because of recurring encounters with the closely spaced walls
within buildings. Locally, the percentage of demolished walls decreases down-current, which has been interpreted as a
loss in kinetic energy within the depositional system of the flow. However, it seems that the upper transport system
by-passed these obstacles, then supplied new pyroclasts to the depositional system that restored its physical
* Corresponding author. Tel.: +39-081-5803114; Fax: +39-081-5527631.
E-mail address: [email protected] (C. Scarpati).
0377-0273 / 03 / $ ^ see front matter < 2003 Elsevier Science B.V. All rights reserved.
doi:10.1016/S0377-0273(03)00146-X
VOLGEO 2630 21-7-03
202
G. Luongo et al. / Journal of Volcanology and Geothermal Research 126 (2003) 201^223
characteristics and restored enough kinetic energy to demolish the next walls and buildings further along its path.
< 2003 Elsevier Science B.V. All rights reserved.
Keywords: Vesuvius; AD 79 eruption; Pompeii; damage to buildings; destructive impact; Polibius; emplacement mechanisms
1. Introduction
The impact of explosive eruptions on densely
inhabited regions is without a doubt one of the
central themes in modern volcanology. The objective of these studies is to develope short- and
long-term strategies aimed at reducing the risks
associated with various styles of explosive eruptions. An accurate evaluation of the volcanic risk
is, nevertheless, not an easily achievable goal, as it
depends not only on the exposed development,
but also on the cumulative knowledge of the
physical behaviour and the past eruptive history
of the volcano.
This paper relates the emplacement of pyroclastic products associated with the well-known Vesuvius AD 79 explosive eruption to the destructive
events in the city of Pompeii; it is aimed at understanding the impact of the eruption on the urban
centres located around the volcano. The study
examines the role of the buildings in a¡ecting
the deposition and resultant distribution of the
pyroclastic products, including a large-scale study
of the e¡ects of the eruption on the city buildings
and an extremely detailed investigation in the C.J.
Polibius’ house, a large mansion located in the
city centre. The studied building was chosen for
the following reasons: (1) the nearby presence of
adjacent outcrops of the pyroclastic deposits of
the AD 79 eruption; (2) the relatively recent age
of the excavation, assuring excellent preservation
of the e¡ects of the eruption; (3) the availability
of excavation reports containing stratigraphic
data and information about the state of the building, including the restoration techniques adopted;
(4) the availability of clear, comprehensive photographic coverage, together with several drawings
and watercolour paintings, documenting the main
phases of the excavation.
This research was performed applying a typical
volcanological approach (stratigraphy and lithofacies de¢nition, reconstruction of the emplace-
ment mechanisms) to the study of these archived
photographs, paintings and documents. This enabled a detailed analysis of the behaviour of these
urban constructions as they were exposed to extreme mechanical and thermal conditions associated with the eruption, as well as evaluated the
extent to which the buildings a¡ected the distribution and dispersion of the pyroclastic products.
In particular, a model for the interaction between
the urban structures and the depositional system
of the pyroclastic density currents (PDCs) is presented. A reconstruction of the sequence of events
in Polibius’ house is also proposed. Finally, a
general model of the e¡ects of the AD 79 eruption
on the city of Pompeii is discussed.
2. The Vesuvius AD 79 eruption
The AD 79 eruption that destroyed Pompeii
and Herculaneum has been widely studied in the
last twenty years and several interpretations of its
dynamics have been proposed (e.g. Lirer et al.,
1973; Sigurdsson et al., 1985; Carey and Sigurdsson, 1987; Cioni et al., 1990, 1992, 1996; Yokoyama and Marturano, 1997; Mastrolorenzo et al.,
2001). The letters, written by Plinius the Younger
to the historian Tacitus, reporting the dramatic
events formed the starting point for all of them,
even the most recent ones.
The eruption was preceded by several earthquakes. The strongest of them, which severely affected the city of Pompeii, occurred in AD 62 and
is considered to be directly related to the AD 79
event (Luongo et al., 1995). The eruption lasted
19 h, from 1.00 p.m. of August 24 to 8.00 a.m. of
August 25. It began with a weak phreatomagmatic explosion covering the volcano with an
ash deposit. Next, a pumice lapilli fall from a
Plinian column 14^32 km high (Carey and Sigurdsson, 1987) covered a large area south of
the volcano (Lirer et al., 1973). Pumices are white
VOLGEO 2630 21-7-03
G. Luongo et al. / Journal of Volcanology and Geothermal Research 126 (2003) 201^223
at the base and grey towards the top of the deposit, re£ecting a change in chemical composition
and density of the tapped magma (Civetta et al.,
1991). The maximum thickness of the fall deposit
(2.8 m) was reached 10 km south of the volcano
in the city of Pompeii (Sigurdsson et al., 1985).
Several PDCs were produced immediately before,
during and after the fallout phase of the grey
pumices; they £owed mainly to the north, west
and south. Cities such as Herculaneum, that
were little, if at all, a¡ected by the pumice lapilli
fallout, were reached by PDCs the emplacement
temperature of which was as high as 400‡C (Kent
et al., 1981; Capasso et al., 2000; Mastrolorenzo
et al., 2001). The development of a syn-eruptive
caldera structure is possibly evident by the pres-
203
ence, only to the volcano slopes, of lithic-enriched, valley ponding £ow deposits (Cioni et
al., 1992, 1996). The role of phreatomagmatism
during the course of the eruption is debated.
The beginning of magma^water interaction is related to the ¢rst PDC after the fall phase (Barberi
et al., 1989; Cioni et al., 1996) or, more probably,
to the last stage of the eruption (Sigurdsson et al.,
1985), as testi¢ed by the abundant accretionary
lapilli and the high degree of fragmentation of
the particles forming the ¢nal strati¢ed ash deposit. The emplacement of the PDCs during this latter stage was not continuous, as is shown by the
presence of three lithic-rich fall layers easily distinguishable in the upper part of the phreatomagmatic sequence.
Fig. 1. Plan view of Pompeii showing the location of the Polibius’ house (grey rectangle), the house of the Chaste Lovers and
their relative outcrops, (a) and (b). Dots represent locations of the other outcrops studied. Roman numbers indicate locations of
‘regiones’. Lower part: facades of the two buildings. Inset: sketch map showing the relative position of Pompeii and Mount Vesuvius.
VOLGEO 2630 21-7-03
204
G. Luongo et al. / Journal of Volcanology and Geothermal Research 126 (2003) 201^223
3. General stratigraphy of the AD 79 deposit at
Pompeii
This stratigraphy is based on seven sections located close to the city walls and inside the city
(Fig. 1). A composite section of the deposit of
the AD 79 eruption cropping out at Pompeii
with a lithological and lithofacies description is
illustrated in Fig. 2. We prefer to use a new descriptive nomenclature rather than the previously
adopted (model-driven) genetic terminology (e.g.
fall, £ow and surge layers by Sigurdsson et al.,
1985 and eruptive units by Cioni et al., 1996). A
correlation between the previous stratigraphic nomenclature of the AD 79 deposit at Pompeii and
that used in this study is also illustrated in Fig. 2.
The AD 79 deposit consists of a lower white to
grey pumice fall bed (units A and B), 2.8 m thick,
overlain by strati¢ed ash and pumice PDC layers
variable in thickness from a few decimetres to 3 m
(units C^L). This upper sequence is interstrati¢ed
with minor, thin and lithic-rich fall layers (units
D, G and I). The main fall bed (unit A) is grainsupported and is composed mainly of angular
pumice lapilli clasts with minor wall rock lithic
clasts of volcanic (lavas) and sedimentary (limestone) origin. Coarse ballistic blocks, both of
pumice and wall rock nature, are scattered
through the bed with a maximum concentration
in the upper grey horizon (unit B). The base of
the PDCs succession is made up of three very thin
layers (unit C): a couple of massive ash layers
with rare, rounded, ¢ne pumice lapilli interlayered
with a very thin sandy layer made up by small,
rounded pumice clasts. This unit is capped by a
massive, 4.5-cm-thick, well sorted fall layer rich in
lithic clasts (unit D). The deposit that overlies this
fall layer is a matrix-supported, strati¢ed bed rich
in rounded pumice lapilli in the basal part with
numerous sedimentary structures (e.g. dunes and
cross strati¢cation) towards the top (unit E). The
uppermost part of the sequence comprises a succession of plane parallel ash layers with abundant
accretionary lapilli (units F, H and L) alternating
with three, very thin, lithic-rich fall layers (units G
and I). The increase of ¢ne ash together with the
presence of accretionary lapilli suggest a change
from magmatic to phreatomagmatic style of the
eruption in correspondence of unit F. Rare impact sags are present at the base of unit G; the
lower ¢ne grained ash layer (F) is deformed by
lithic fragments up to 9 cm in diameter.
The main structural features of the recognised
units are exempli¢ed at the Nola Gate section,
where the AD 79 pyroclastic sequence thickens
in correspondence of a Roman road enclosed between an embankment that rests on the city walls
and an enclosure wall (Fig. 3). Both the thick
coarse pumice fall layers (units A and B) and
the thin lithic-rich fall layers (units D, G and I)
maintain a uniform thickness on the steep slope
(25‡) of the Roman embankment. The lower PDC
unit (C) shows a relatively uniform thickness on
the slopes (4^8 cm) and thickens in the depression
(25 cm). The main PDC unit (E) is topographically controlled, ¢lling the depression. Its lower
layers are completely con¢ned within the depression (E1), while its upper layers drape the topography thinning toward both the sides (E2).
4. Stratigraphy inside Pompeii
The most complete stratigraphic sections of the
AD 79 deposit inside the city are located a few
metres from the eastern wall of the Polibius’
house, in the garden and immediately west of
the house of the Chaste Lovers, which is, in
turn, located a few metres west of the Polibius’
house, i.e. locations (a) and (b) in Fig. 1. The
detailed stratigraphic reconstruction of these two
sections is of absolute importance, even beyond
the objective of the present study, since they represent extremely rare examples of deposits of the
AD 79 eruption still preserved in the city of Pompeii. Most of the previously studied stratigraphic
sequences were, indeed, located either outside the
city walls (Sigurdsson et al., 1985; Santacroce,
1987; Lirer et al., 1993) or in their vicinity (Rittmann, 1950; Di Girolamo, 1964; Cioni et al.,
1996; Yokoyama and Marturano, 1997).
4.1. The house of the Chaste Lovers
This outcrop is located in a narrow alley (a few
metres wide) between two vertical walls on the
VOLGEO 2630 21-7-03
G. Luongo et al. / Journal of Volcanology and Geothermal Research 126 (2003) 201^223
VOLGEO 2630 21-7-03
Fig. 2. Left: composite stratigraphic section showing the maximum thickness of the units of the AD 79 deposit at Pompeii. A lithofacies description is also illustrated. Right: detailed stratigraphic columns along the western side of the house of the Chaste Lovers and in a hole dug a few metres from the eastern perimeter
wall of the Polibius’ house. For location of the sections, see Figs. 1 and 4. Inset: relationship between the stratigraphy presented in this paper (TP) and those published by Sigurdsson et al., 1985 (S) and Cioni et al., 1996 (C).
205
206
G. Luongo et al. / Journal of Volcanology and Geothermal Research 126 (2003) 201^223
Fig. 3. (a) The complete sequence of the AD 79 pyroclastic deposit at Nola Gate (for location, see Fig. 1). The pyroclastic deposit thickens in correspondence of a Roman road enclosed between an embankment that rests on the city walls and an enclosure wall. (b) Sketch of the geometry of the di¡erent units distinguished throughout the AD 79 deposit. Note the mantling nature
of the fall units (A, B, D, G, and I) while the PDC units show di¡erent geometric relationships with respect to the palaeo-topography. Unit C mantles the topography and thickens in the depression; unit E is valley-ponding, ¢lling the topographic depression. The upper part of the sequence (dotted layer) is made of reworked material accumulated during past excavations (actually
this material has been removed).
western side of the house of the Chaste Lovers
(Reg. IX, 12, 5^7). The outcrop is 28 m long
and averages 2.5 m in height. The cut strikes
N150‡ and is radially oriented with respect to
the vent. The succession is made up of a PDCs
deposit of strati¢ed ash and pumice lapilli overlying the thick pumice lapilli fall (Fig. 4); the base
of the deposit (white pumice lapilli, unit A) is not
outcropping. Fig. 2 shows ¢ve stratigraphic sections that have been measured in detail from the
outcrop and lateral facies variations. The fall de-
posit (unit B) is not continuously outcropping and
its thickness ranges from 0 to 1.2 m. It is massive,
composed of grey pumice lapilli and minor lithic
clasts (both lavas and limestone) showing a maximum diameter of 7.5 cm. An easily recognisable
erosional surface separates it from a 8-cm-thick
unit made up of ¢ne, sub-rounded pumice clasts
with increasing ¢ne ash toward the top (unit C). It
is overlain by a massive, well sorted, 6-cm-thick
unit formed of angular lithics and minor pumice
clasts (unit D). This whole sequence is covered by
VOLGEO 2630 21-7-03
G. Luongo et al. / Journal of Volcanology and Geothermal Research 126 (2003) 201^223
207
Fig. 4. Sedimentary structures in the AD 79 deposits close to the house of the Chaste Lovers. (a) Truncated walls of the house
of the Chaste Lovers (foreground); the AD 79 pyroclastic succession (background). (b) Sketch of the stratigraphic succession
showing the main sedimentary structures of the deposits. The wave-shaped boundary line between the basal fall pumice and the
overlying PDCs deposit is worth noting. The progressive crests of the dunes are well evident in the middle part of the sequence.
The arrows show the exact location of the detailed stratigraphic sections reported in Fig. 2. The scale changes in this perspective.
a 69^115-cm-thick bed showing dunes the average
wavelength and amplitude of which are 15 and
0.7 m, respectively (unit E1); they are formed of
alternating layers of ¢ne ash and coarse pumice
lapilli becoming thicker and coarser in down-current direction. The dune pro¢le is symmetric, both
the £anks of the crest dipping around 15‡. If compared to subaqueous analogues, these dunes can
be classi¢ed as ‘medium dunes’ (Ramsay et al.,
1996). Another, smaller kind of dune that is ¢ner
grained, is present in between the crests of the
previous dune system. Their wavelength/amplitude ratio is lower (6 vs. 20) and their maximum
thickness reaches 35 cm (unit E2; Fig. 5). They
are comprised of alternating layers of ¢ne and
coarse ash; their crests are cuspate and asymmetrical, the up-current and the down-current £anks
dipping 15‡ and 13‡, respectively. The lee slope is
also concave upward. These dunes are progressive
since their crests tend to migrate down-current
towards the southeast. Massive and accretionary
lapilli-rich ash layers, 16^45 cm thick, overlie the
VOLGEO 2630 21-7-03
208
G. Luongo et al. / Journal of Volcanology and Geothermal Research 126 (2003) 201^223
Fig. 5. Unit E2 forming progressive dunes in the central part of the pyroclastic current deposits at the house of the Chaste Lovers. Flow direction is from right to left. Scale is in 20-cm intervals
dune beds (unit F). Each bed shows an upward
increase of the abundance of accretionary lapilli.
Above them, two lithic-rich fall levels form a
characteristic couple separated by a 2-cm-thick
ash layer (unit G). The lithic-rich levels are massive, well sorted and made up by angular clasts;
their respective thicknesses are 4 and 3.5 cm. At
the top of this unit another ash deposit with accretionary lapilli, 12^15 cm thick, is present (unit
H). Overlying this unit there is one more thin (1.5
cm), lithic-rich fall level (unit I). Finally, a strati¢ed ash deposit, 20^62 cm thick, caps the sequence (unit L): individual beds are massive and
contain sparse accretionary lapilli. The whole succession is locally eroded and channels are ¢lled
with reworked pyroclastic material, 10^90 cm
thick (unit M).
4.2. Near the Polibius’ house
This outcrop is located on the eastern side of
the room [DD] in Polibius’ house; some further
excavation was required to obtain better exposure. The section is almost 2 m high, but the
base of the deposit cannot be observed. The pum-
ice lapilli layer (unit B), 83 cm thick, is overlain
by a 4-cm-thick ash horizon that includes a thin
(1 cm) coarse ash unit (unit C). A massive, well
sorted, 5-cm-thick fall bed, mainly formed of
lithics and minor pumice clasts (unit D), follows.
This part of the sequence ends up with a lens of
reworked material cut into the underlying units
up to the top of layer B. Above it, the deposit
is strongly layered and mainly made up of
PDCs products. The ¢rst horizon is made up by
well rounded pumiceous lapilli and subordinate
ash (unit E1); the lapilli are mainly concentrated
in the lower part (thickness ranges from 19 to 38
cm). Next, a conformable layered ash horizon,
9^16 cm in thickness, is present (unit E2). This
level is, in turn, overlain by another ash level,
11 cm thick, massive and very rich in accretionary
lapilli (unit F). Two thin lithic-rich and well
sorted airfall layers, respectively 3 and 1.5 cm,
overlie unit F. They enclose a 1.5-cm-thin ash horizon (unit G). The ¢nal deposit is a layered ash
horizon, 56 cm thick (unit H and L), with a thin
lithic-rich level (unit I) located a few centimetres
from the base. The sequence is covered by the
modern soil.
VOLGEO 2630 21-7-03
G. Luongo et al. / Journal of Volcanology and Geothermal Research 126 (2003) 201^223
5. Investigation in the C.J. Polibius’ house
5.1. Location and description of C.J. Polibius’
house
C.J. Polibius’ house is located in Via dell’Abbondanza (Reg. IX, 13, 1^3), one of the main
roads in the centre of Pompeii (Fig. 1), at a height
of 32 m above sea level; the ground on which it is
located slopes towards the road. The shape of the
building is nearly rectangular, 48 by 18 m, with
209
the larger side perpendicular to the main road; its
area is 850 m2 (Fig. 6). The building is bounded
by roads on the southern and western sides and
by other buildings to the north and east. Two
£oors are present: 37 rooms are located on the
ground £oor and 19 on the upper £oor, the former being identi¢ed by means of one or two capital letter(s) in Fig. 6.
For the sake of simplicity, the house can be
subdivided into three zones: front, middle and
rear. In the front part two £oors are present:
Fig. 6. Ground £oor plan view of Polibius’ house. Rooms indicated by capital letters. Rooms of which roofs and walls collapsed
during the ¢rst fall phase are distinguished from those of which the building structures collapsed during the emplacement of
PDCs. Note the directions of upsetting of the walls during the two main eruptive phases: random directions in the fall deposit
and exclusively towards south in the PDCs deposit. The locations of the stratigraphic sections illustrated in Figs. 7 and 11 are
also shown.
VOLGEO 2630 21-7-03
210
G. Luongo et al. / Journal of Volcanology and Geothermal Research 126 (2003) 201^223
Fig. 7. Stratigraphic sections inside the Polibius’ house showing the pyroclastic succession de¢ned with lithostratigraphic data
from the excavation reports. The thickness of the distinguished deposits (fall lapilli pumice deposit, strati¢ed ash and pumice
PDCs deposit, reworked pumices and soil) change remarkably in the di¡erent rooms. The external pro¢le of the reconstructed
building is also reported. Zero height corresponds to the ground level at the time of the excavation; horizontal distance is not to
scale. Section traces are reported in Fig. 6.
VOLGEO 2630 21-7-03
G. Luongo et al. / Journal of Volcanology and Geothermal Research 126 (2003) 201^223
the maximum height is about 9 m along Via
dell’Abbondanza and 6.5 m towards the interior.
Two impluvia are present in this area, a main one
[O] and a secondary one [N]. The middle zone of
the building is a square-shaped garden 100 m2
wide; this is surrounded by a portico on three
sides (north, east and south) and bounded by
the 5.4-m-tall perimeter wall of the building on
the fourth side. In the rear zone the upper £oor
is absent and the maximum height reached is
6.2 m. The roofs are very articulated ; according
to the current reconstruction of the house, none
of them were £at. In the front zone of the house,
they were re-built in order to almost completely
drain rain water towards the impluvia [O] and
[N]. The roo¢ngs on the northern, eastern and
southern porticos dip towards the garden; the
one on the western side bears a cuspate roo¢ng
dipping towards both the garden and the outer
road. Because of the sloping ground towards
Via dell’Abbondanza, the position of the £oor
rises toward the rear part of the building. As a
general rule, the rooms in the front are either at
the same level of [A] or 70 cm higher; in the
garden and the rear zone the £oor elevation is
1.2 m.
5.2. Stratigraphy of the AD 79 deposits inside the
Polibius’ house
Based on the large number of observations reported in the excavation logs, a reliable stratigraphic reconstruction could be made in the
majority of the Polibius’ house rooms. The stratigraphic sequence in 29 of the 37 rooms on the
ground £oor was reconstructed. Fig. 6 illustrates
the rooms for which at least one datum is available. In order to synthetically show the stratigraphic sequence, eight sections were reported
(Fig. 7) showing: (1) the £oor pro¢le; (2) the
roof pro¢le and the location of the outer perimeter walls of the house according to the way they
were rebuilt; (3) the lithotypes observed at the
di¡erent stratigraphic heights ; (4) the position of
the boundary between the white^grey lapilli fall
deposit and the upper strati¢ed ash and pumice
PDCs deposit; and (5) the upper limit of the
strati¢ed ash and pumice PDCs deposit.
211
With respect to its main features, this sequence
looks very similar to those previously described in
Sections 4.1 and 4.2: a thick basal pumice deposit
is overlain by a PDCs succession locally covered
with reworked material or humus. Nevertheless,
the thickness variations observed in the pumice
deposit (1^5 m) are very signi¢cant: the greatest
thicknesses are observed in the two impluvia [N]
and [O], in the peristyle and the alleys and gardens located immediately outside the house (sections 2, 3 and 5; Fig. 7). The overlying PDCs
succession also shows thickness variations (from
few tens of cm to 4 m). In this case the deposit is
thicker where the pumice lapilli fall is thinner and
vice versa ; an approximately uniform total thickness is thus maintained. The reworked material,
the thickness of which is also variable, is sometimes reported as being incorporated into the primary deposit. This is a case of the reworking taking place during the breaching of walls when the
¢rst exploratory wormholes were dug. In other
cases, i.e. [II] and [HH], the primary deposit was
almost totally lacking and only reworked material
was found. The whole sequence is capped by a
soil whose thickness ranges from a few tens of
centimetres to 2 m. The presence of a peculiar
stratigraphic sequence (section 3 and Fig. 7) in
[II], [GG] and [HH] is extremely interesting.
Here, an ash layer of a few centimetres to 40^50
cm thick is directly laid on the £oor. Above it, a
thick pumiceous deposit probably of reworked
material is present. Along the western wall (still
intact) of the peristyle, dune-shaped bedforms related to the second phase of the eruption are
easily recognised (Fig. 8): their wavelength is
around 8 m and their amplitude ranges from 30
cm for the lowermost to 45 cm for the uppermost
one. Their crests do not show any clear up- or
down-current migration, thus implying a stationary wave. The height of the base of the lowermost
dune marks exactly corresponds to the boundary
between pumice fall and PDCs deposit as described in the excavation books.
5.3. State of the objects found
A remarkable number of objects were found
during the excavations in the building. A thor-
VOLGEO 2630 21-7-03
212
G. Luongo et al. / Journal of Volcanology and Geothermal Research 126 (2003) 201^223
Fig. 8. Left: picture of the western wall of the peristyle. The marks of two wavy surfaces can be observed above and below the
roo¢ng testifying to the presence of dunes from a typical pyroclastic current deposit. Right: reconstruction of the boundary between the fall pumice and the strati¢ed PDCs deposit. The marks left by the pyroclastic currents deposit in the western wall of
the peristyle are reported: it can be observed that the basal mark corresponds to the boundary between the two deposits.
ough analysis of all of them would be far beyond
the aim of the present research. Rather, it was
decided to focus our attention on those fragile
objects such as glasses and earthenware, the condition of which was a good indicator of the thermal and mechanical stresses during the di¡erent
eruptive phases and the burial of the house. Several amphorae crushed by a big fallen roof were
found on the £oor of [A], amphorae and other
earthenware were found intact on the £oors of
[TT], [C], [UU] and [BB]: all of them were enclosed in a thick layer of pumice lapilli. Many
other objects, intact or broken, were found in
the pumice lapilli deposit 1^3 m above the respective £oors in rooms [D], [N], [G], [A], [R], and
[BB]. Earthenware, mainly in fragments, were
also found in the PDCs succession. The objects
contained in 5 closets located under the roof
along the eastern wall of the peristyle formed an
extremely interesting discovery. The lower part of
the closets was buried under the pumice lapilli
and the upper part under the ash and pumice
PDCs deposit. The wood slowly rotted out during
the centuries leaving several astonishingly well
preserved objects in the hardened ash and pumices: bottles, jugs, dishes, jars, glasses and several
earthenware pieces. Most of them were intact and
perfectly preserved.
5.4. Carbonised wood and traces of ¢re
Remains of carbonised wood were found in the
front part of the house. They mainly consisted of
roofbeams supporting the ceiling of the ground
£oor and, to a minor extent, tables or doorposts,
still in place or collapsed to the ground. All of
these remains were found preferentially within
the fall pumice and only rarely within the PDCs
deposit. The condition of the plasters and frescoes
on the walls of the ground £oor ruled out the
possibility that great ¢res took place during or
immediately after the eruption. Only the walls of
room [Z] were entirely covered in soot, probably
as a result of local combustion. In addition, some
studies carried out on the glasses found in the
building revealed that only in one case (bottleneck
from closet IV in the peristyle; Verita' et al., 2001)
the temperature reached a value su⁄ciently high
(620‡C) to completely deform the object.
5.5. Skeletons
All of the recovered skeletons were found in
rooms [GG] and [HH] in the rear of the building,
the only exception being two human skulls coming from the upper part of the pumice lapilli deposit (probably locally reworked) in the peristyle.
VOLGEO 2630 21-7-03
G. Luongo et al. / Journal of Volcanology and Geothermal Research 126 (2003) 201^223
213
Fig. 9. Some of the skeletons in room [GG]. Skeletons (III) and (IV) are above an ash layer, 20^40 cm thick, lying directly on
the £oor. The skull of skeleton (I) is close to the feet of skeleton (IV).
All of the four skeletons from [GG] (Fig. 9)
were lying on an ash layer variable in thickness.
The head of the ¢rst (I) skeleton was 30 cm above
the £oor in the northeastern corner; the remaining part of the body was found in the northwestern corner, partly directly on the £oor and partly
30 cm above it. The second (II) and third (III)
skeletons were lying on the £oor, respectively
along the western wall and in the middle of the
room. The fourth (IV) skeleton was found along
the eastern wall; the legs and the rest of the body
were respectively 20 and 40 cm above the £oor.
Seven skeletons were recovered in [HH]. The
¢rst (I) skeleton, located along the northern
wall, was lying above some tiles and other ceiling
debris ; it was buried into a mixed deposit of ash,
pumices and stucco fragments from the ceiling.
The second (II) and third (III) skeletons were on
the £oor in the northwestern corner; they were
covered with pumice lapilli and ceiling stucco.
The fourth (IV) skeleton was found along the
western wall, 40 cm above the £oor. The ¢fth
(V) and sixth (VI) skeletons were in the middle
of the room, lying on a thin ash layer and under
stucco fragments from the ceiling. The seventh
(VII) skeleton consists of the small bones of a
foetus found in the abdominal region of the
mother’s skeleton (II). Some casts were found
within the ash deposit. They are related to some
tables, probably beds, located along the northern
and western wall in [HH]. Skeletons (I) and (IV)
may originally have been lying upon them.
5.6. Collapses
A great amount of debris from walls and roofs
was found. According to the excavation reports,
collapses during the eruption lowered the average
height of the walls 3^4 m.
5.6.1. Attics and roo¢ngs
At the moment of excavation, no roofs or attics
were found to be in situ; debris from them was
found both in the basal pumice fall and the over-
VOLGEO 2630 21-7-03
214
G. Luongo et al. / Journal of Volcanology and Geothermal Research 126 (2003) 201^223
lying strati¢ed ash and pumice PDCs deposit (Fig.
6). Parts of the roof of [A], [M], [E], and [D], the
attic of [Z], and the roo¢ngs of [N] and [G], all of
them recovered from the pumice lapilli, were lying
on the respective ground £oors; the attics of the
upper £oors and the roofs of [C], [B], [H], [SS],
[AA], [GG], [II], [HH], and [BB] were found in the
PDCs deposit. In particular, the ceiling of [SS]
was found to be only 20 cm lower than its original
level. On the contrary, the western, eastern and
northern roo¢ngs of the peristyle, buried by the
strati¢ed ash and pumice deposit, were found still
above the perimeter wall. The roo¢ng on the
northern side was partially destroyed, but the
others su¡ered very little damage. Only few roofing-tiles and bent-tiles were missing from the
western roo¢ng, whereas the eastern one was
practically still intact. It is important to mention
that the original level of these roo¢ngs was 2 m
lower than the level of those in the front and rear
zones of the house. The stratigraphic height of the
tiles within the volcanic products in [EE] establishes the timing of collapses of the ceiling in the
building. In the western part of the room the ceiling was smashed to the ground under the load of
the pumice lapilli fall deposit, whereas in the eastern part the cast of a bed headboard bent by the
falling ceiling was observed within the PDCs deposit. In addition to the roo¢ngs and attics described above, several tiles, either intact or in
fragments, were recovered at various stratigraphic
heights in all the rooms of the building. For instance, fragments of tiles were recovered from
throughout the whole thickness of the deposit in
room [EE].
5.6.2. Walls
Nearly all of the walls were found to be partly
or entirely demolished. In particular, only the
western corner of the facade and the lower part
of some inner walls were still in place on the
upper £oor. According to the excavation reports,
the walls buried in the PDCs deposit had mainly
collapsed southwards, whereas those in the fall
lapilli did not show any predominant direction
of collapse (Fig. 6). It is also worth mentioning
that, as a general rule, the north^south trending
walls were less damaged than those trending east^
west. Fig. 10 shows the average heights of the
transversal walls of the house. Even removing
the topographic e¡ects (sloping ground), the
amount of demolition was higher moving from
the main facade towards the peristyle and then
decreases again in the rear of the building. Fig.
11 shows ¢ve transversal cross-sections along
walls located at increasing distance from the facade of the building. In all of them, a channelshaped pro¢le of the demolished walls can be observed. These pro¢les have di¡erent depths and
their bases have been measured at di¡erent
heights from the ground £oor. The northernmost
pro¢le is reported in the excavation reports at the
top of room [EE] (Fig. 11), with its depth less
then 1 m and its base at the top of the ground
£oor (about 6 m height). The middle pro¢le is
more then 6 m deep and its base is a few tens
of centimetres above the ground £oor. The southernmost one is about 4 m deep and its base is
nearly on the upper £oor. These pro¢les largely
correspond to the boundary between the basal fall
and the overlying PDCs deposit. Within the house
Fig. 10. Variations in the average height of the east^west trending (transversal) walls in the Polibius’ house; the variations of the
ground level are taken into account.
VOLGEO 2630 21-7-03
G. Luongo et al. / Journal of Volcanology and Geothermal Research 126 (2003) 201^223
215
Fig. 11. Sections along some of the transversal walls of the house showing the concave upper pro¢le of the walls. (r) Top of
room [EE] redrawn from excavation journals. (V), (VI) and (VIII) sections reconstructed during this study. (f) Picture of the facade as found during the excavations. Boundaries between the di¡erent lithotypes are shown for sections (V), (VI) and (VIII).
The correspondence between the wall heights and the contact between the fall and the PDCs deposit is worth noting. The locations of the sections are reported in Fig. 6. Legend as in Fig. 7.
VOLGEO 2630 21-7-03
216
G. Luongo et al. / Journal of Volcanology and Geothermal Research 126 (2003) 201^223
it is important to point out two other intriguing
observations: (1) the transverse walls in the rear
of the house were still intact at the time of excavation; and (2) the metal hinges of the doors of
the rear rooms lay on the £oor under the PDCs
deposit.
5.7. Extent to which the roofs and walls a¡ected
the distribution of the pyroclastic fall deposit
The stratigraphy of the products of the AD 79
eruption inside the Polibius’ house is anomalous
with respect to the surrounding areas (Fig. 7), the
main di¡erence consisting in thickness changes of
the fall deposit. Remarkable variations were indeed observed in the peristyle [CC], in the alleys
around the building, and, as already reported in
Section 5.2, in rooms [N] and [O]. Also, the fall
pumice deposit in the garden of the peristyle thins
towards the sheltered corridor of the eastern portico (section 4; Fig. 7), probably as a result of
pumices rolling down from the thickening heap
in the centre of the garden.
The present restoration of the building, with all
of the roofs dipping at 15‡ towards both the surrounding alleys and rooms [CC], [N] and [O],
would explain the additional amount of fallout
deposit as being due to the rolling down of the
falling pumices from the adjacent roofs.
Evaluating the exact timing of the roof collapses is not straightforward. If an accumulation rate
of 15 cm/h (Sigurdsson et al., 1985) and an average density of 650 kg/m3 are assumed for the falling pumice, an incremental load of 100 kg/m2 per
hour is realised. Unfortunately, since no data on
the yield strength of the Pompeii roofs are available in the literature, the critical thickness causing
the collapse cannot be determined. On the other
hand, studies carried out on several recent eruptive events (Blong, 1984) report roof collapses to
occur with thicknesses ranging 0.1^1 m, a value
well matched by the 40 cm calculated by Sigurdsson et al. (1985). As a consequence, roof collapses
in Polibius’ house might have taken place 1^6 h
after the beginning of the eruption. On the other
hand, the PDCs deposit belonging to the second
eruptive phase, lying directly on the £oor in many
rooms of the rear part of the house, suggests some
of the roofs be still in place after the end of the
initial pumice fall phase. All of the above data
clearly show that a di¡erent response of the roofs
of the building took place during the ¢rst phase of
the eruption. Those in the front part of the house
collapsed, whereas those in the central and rear
zones did not. Since the possibility of di¡erent
pumice amounts falling on the di¡erent parts of
the building can be ruled out completely, it is
proposed that the roofs in the front zone were
much £atter than those in the current restoration
of the building and allowed an unusual accumulation of the falling pumices, eventually collapsing
under the pumice load. The steeper roofs in the
rear part of the house and the peristyle allowed
the falling pumices to roll and slide down and
accumulate in the adjacent impluvia and the garden, respectively.
6. State of the buildings in the whole city of
Pompeii
To investigate the damage su¡ered by the
whole city of Pompeii we evaluated the state of
the buildings all over the city (Fig. 12a). A limit
to this investigation is the partial reconstruction
(restoration) of many buildings and the lack of
good reports from ancient excavations (most of
the city has been excavated in the past centuries).
Our survey shows that the type of destruction
su¡ered by the city can be described in the following observations:
(1) The ground £oor is partly intact in most of
the buildings, whereas the upper £oor is almost
completely demolished.
(2) Most of the destruction is stratigraphically
related to unit E. In fact, as exempli¢ed by the
recently excavated House of Chaste Lovers, material from demolished walls lay inside this unit
(Fig. 12b,c) as well as some of the victims of
the PDCs phase of the eruption (Luongo et al.,
2003).
(3) The east^west oriented walls are by far more
damaged than those striking north^south.
(4) It is still possible to observe a channelshaped pro¢le in many of these demolished walls.
(5) In many cases, the northern, vent-facing
VOLGEO 2630 21-7-03
G. Luongo et al. / Journal of Volcanology and Geothermal Research 126 (2003) 201^223
217
Fig. 12. State of the buildings at Pompeii. (a) Aerial view of the city. Note that the upper £oors of most buildings are missing.
(b) Front view of a wall pulled down in the house of the Chaste Lovers. The wall, transversal to the £ow direction, lays in the
E1 unit. Note the darker, lithic-rich unit D below the wall. (c) Rear view of the same toppled wall. On the left, the still standing
wall found in the fallout deposit; on the right, the upper part of the wall laying above unit D. (d) Section of the city walls close
to Capua Gate showing a high amount of destruction in the northern, volcanic vent-facing side. Flow direction from right to
left. Key: FU, fall units; PDCs U, pyroclastic density current units. The arrow links the base of the PDCs deposit with the pro¢le of the destroyed wall.
part of the buildings was more damaged than the
southern one (Fig. 12d).
(6) The northern (relatively proximal) and
southern (relatively distal) sectors in the city
were a¡ected in the same way.
7. Analysis of the behaviour of the pyroclastic
currents
A major contribution to understanding the
PDCs transport and emplacement mechanisms
was provided by comparing large-scale (i.e. whole
city) and small-scale (i.e. building by building)
distribution of damages. The distribution and
type of the damages noticed throughout the city
is inhomogeneous although the east^west oriented
walls are more completely destroyed. Locally, a
more regular destruction pattern is recognisable;
the height of demolished walls decreases toward
south.
Recent studies (Valentine, 1987; Fisher, 1990;
Druitt, 1992, 1998; Freundt and Bursik, 1998;
Freundt et al., 2000) and direct observations of
PDCs (e.g. Mount St. Helens and Pinatubo eruptions) showed that these currents are clouds of gas
and volcanic ash tens to hundreds of metres thick
thee particle concentration of which increases towards the base. In these £ows particles are mainly
transported in suspension by turbulence (trans-
VOLGEO 2630 21-7-03
218
G. Luongo et al. / Journal of Volcanology and Geothermal Research 126 (2003) 201^223
port system) and settling progressively from a
basal depositional system. During the transport
a vertical £ux of grains (suspended-load fallout)
is supplied from the transport system to the
underlying depositional system. Deposition is
controlled by particle concentration, suspendedload fallout rate, the grain-size distribution and
the £ow power (Lowe, 1982, 1988; Druitt,
1991). The presence of di¡use sedimentary structures within the PDCs AD 79 deposit at Pompeii
suggests the dilute and expanded nature of these
pyroclastic currents. As a consequence, only the
lowermost part of such currents has interacted
with the buildings at Pompeii.
To explain type and distribution of the damage
mentioned above, the following model of the emplacement of the AD 79 PDCs at Pompeii is proposed. The ¢nding of toppled walls (and victims)
in unit E indicates a cause^e¡ect relationship between the emplacement of the parental PDC and
the destruction of the buildings. This means that
the thin ash layer C was emplaced without any
signi¢cant amount of destruction. Probably, it
was deposited by the distal and dilute part of a
PDC. In fact, a decrease in concentration away
from the vent has been inferred from ¢eld features
of several PDCs deposits (e.g. the 18 May 1980
Mount St. Helens blast, Druitt, 1992; the Suwolbong tu¡ ring in Korea, Sohn and Chough, 1989).
The large-scale inhomogeneous distribution of
damage suggests a non-uniform behaviour of the
pyroclastic currents. As in the 1902 St. Pierre
eruption surges (Fisher and Heiken, 1982), the
AD 79 pyroclastic currents demolished mostly
the walls that happened to be perpendicular to
their £ow direction. The ¢nding of well preserved
Fig. 13. Sketch of a PDC £owing on the Pompeii urban structures. It is possible to distinguish an upper transport system that
supplies, with an uniform rate, particles to a basal depositional system, variable in thickness. Two series of walls, transversal to
the £ow direction, are widely spaced. Near the ¢rst two obstacles (walls on the left), the depositional system is thin with a relatively low concentration of pyroclasts. From the ¢rst to the second series of obstacles, there is enough space to restore a thick
basal layer with relatively high concentration. This basal part has su⁄cient energy to deeply topple the transversal walls. The
presence of the following closely spaced obstacles slow down the depositional system that losses part of its load; as a consequence, the thickness and concentration of the depositional system decreases progressively in correspondence of each wall.
VOLGEO 2630 21-7-03
G. Luongo et al. / Journal of Volcanology and Geothermal Research 126 (2003) 201^223
brittle objects in closets located laterally to the
channel-shaped pro¢le of the demolished walls,
inside the Polibius’ house, testi¢es that the destructive impact of the PDCs is restricted along
well de¢ned paths. Indeed, the areas outside the
channels represent places where PDCs show essentially depositional behaviour. The channelling
of the basal part of the currents along the central
axis of the buildings is probably an e¡ect of the
bounding of the longitudinal walls. Furthermore,
as in the case of the Polibius’ house, the walls were
progressively less damaged down-current for each
single building; however, where an open space
(10^20 m) occurs between two closely spaced
(2^3 m) series of walls, the amount of demolition
became high again (Fig. 13). It is suggested that
this trend could re£ect a temporary reduction of
kinetic energy within the depositional system due
to local loss of mass (sedimentation) and velocity
after a series of repeated impacts against the
buildings. The loss of kinetic energy reduces the
capacity of the PDCs to destroy the walls and
erode the underlying pumice fall deposit. The
basal depositional system could restore its physical characteristics once every series of obstacles
was overridden thanks to the continuous supply
of particles from the upper high-energy transport
system that did not su¡er any breaking action by
these urban obstacles. A new decrease of kinetic
energy occurs wherever a new series of closely
spaced obstacles occurs transversal to the £ow direction. The depositional system is non-uniform in
terms of concentration and occurs as a series of
longitudinal rarefactions and condensations in response to the distance between successive obstacles (assuming a constant supply of particles
from the upper transport system). Consequently,
we suggest that local and temporary £uctuations
of kinetic energy within the depositional system
can be induced by the interaction between the urban structures and the base of the PDCs.
Recently, Valentine (1998) estimated the dynamic overpressure necessary to completely destroy the buildings of the nearby city of Herculaneum by the PDCs of the AD 79 eruption at 35^
70 kPa. The extensive damage su¡ered by the
Pompeii buildings is possibly related to an analogous PDC overpressure.
219
There is peculiar damage in the Polibius’ house
that allows us to re¢ne the proposed £ow model.
We know both from the volcanic vent location
and the sedimentary structures ^ namely dunes
and cross-strati¢cation ^ that the pyroclastic currents £owed southwards. The location of the
hinges inside the rear rooms of the Polibius’ house
demonstrates the knocking down of the doors
towards the north (contrary to the previously
documented north^south PDCs £ow direction).
Laboratory experiments about the e¡ects of topography on propagation and sedimentation of
density currents con¢rm that the presence of an
obstacle of su⁄cient amplitude and transversal to
the £ow direction will lead to a partial blocking of
the current and the production of a wave which
propagates upstream (Bursik and Woods, 2000).
We suggest that the present evidence is consistent
with the production of a re£ected £ow against the
southern transversal walls of the peristyle. This
re£ected £ow moved across the peristyle towards
the rear of the house that was sheltered by the
direct impact of the PDCs by the adjacent buildings located to the north of the Polibius’ house. It
struck the wall and knocked down the doors towards the interior.
8. Emplacement temperature of the pyroclastic
products
Several lines of evidence rule out the possibility
of a high emplacement temperature for all of the
volcanic products covering Pompeii. In the ¢rst
place, the typical textural features of high-temperature deposits (Branney and Kokelaar, 1992;
Freundt and Schmincke, 1995) are totally absent
from the observed stratigraphic sections of the
AD 79 eruption. Moreover, a large amount of
accretionary lapilli are present in the upper part
of the stratigraphic sequence (units F, H and L),
testifying to a depositional temperature not in excess of 100‡C (Schumacher and Schmincke, 1991).
Other strong evidence is provided by the large
number of fugitive bodies buried in the middle
of unit E (Luongo et al., 2003): the temperature
of the previously accumulated products (of both
fall and PDC origin) would not have been so high
VOLGEO 2630 21-7-03
220
G. Luongo et al. / Journal of Volcanology and Geothermal Research 126 (2003) 201^223
as to prevent the Pompeii inhabitants from attempting to escape the disaster. Further evidence
comes from studies carried out on the previously
mentioned (Section 5.4) deformed glass fragment,
the present tensional state of which results from
rapid cooling (Verita' et al., 2001), thus ruling out
the presence of a very hot and slowly cooling enveloping pyroclastic deposit. Initially, the glass
probably softened because of a local combustion
in the cupboard. In the same way, the large
amount of carbonised wooden relics recovered
from the pumice lapilli are to be attributed to
mineralisation rather than to combustion phenomena.
9. Chronology of the events: the e¡ects of the
AD 79 eruption on Pompeii
A reconstruction of the timing of the destructive events in the city has been attempted. In the
early afternoon of August 24, a pumice fall
started burying the city of Pompeii. During this
¢rst phase of the eruption several roof collapses
occurred, evident by abundant debris and tiles
found in the lapilli fall deposit. A signi¢cant number of victims has also been found in this deposit
(38% of the total of Pompeii victims; Luongo et
al., 2003). They probably died as a consequence
of the building’s collapse. The accumulation rate
of the deposit in the outdoor areas of the city is
estimated at 15 cm/h. In the zones of the buildings
where material sloughed from roofs, the accumulation rate reached values as high as 25^30 cm/h.
Within 6 h the roofs and part of the walls of the
buildings had collapsed under the pumice load.
On the morning of August 25, many buildings
had been destroyed almost completely and only
some edi¢ces were still standing. A remarkable
thickness of pumice lapilli, 1^5 m, had ¢lled the
impluvia and the rooms the roofs of which had
collapsed. Also, the peristyles and the alleys had
been ¢lled to these depths. The lapilli pumice fall
deposit, generally 3 m thick, totally buried the
lower part of the buildings, partly protecting
them from the impact of the later pyroclastic currents. The ¢rst PDC £owed through the city depositing the basal ash layer C and causing irrele-
vant damage. A discrete explosion produced a fall
that formed the lithic-rich layer D. A local episode of reworking suggests a brief pause before
the emplacement of the next PDC. It showed a
greater destructive force, £attened most of the
(especially transversal) walls in its north^south
path. This PDC buried the ruins under a few
metres of coarse ash and pumice deposit (unit
E). A ¢nal phreatomagmatic phase, punctuated
by three lithic fall episodes, emplaced the upper
part of the succession (F^L units).
A more detailed scenario has been reconstructed for the Polibius’ house. During the ¢rst
phase of the eruption a great quantity of pumice
lapilli fell on the building (Fig. 14a), producing
the collapse of the front part of the house (Fig.
14b) in a few hours. Probably as a consequence of
these events, the inhabitants of the Polibius’ house
decided to hide in the rear rooms [GG] and [HH],
the roofs of which, probably steeper than those in
the front of the house, had not been damaged by
the falling material. The pyroclastic currents arrived from the north, travelling parallel to the
long sides of the building: the northernmost
wall, east^west oriented, was little, if at all, damaged by these currents because of the protection
by the adjacent buildings. The pyroclastic currents
overtopped the rear part of the house. They
moved into the garden and advanced towards
the front of the house. We suggest that the PDC
that emplaced unit E eroded the underlying pumice fall deposit and totally destroyed the upper
£oor and the ground £oor walls perpendicular
to the £ow direction (Fig. 14c). This current was
channeled between the longitudinal walls focusing
its destructive power along the axis of the channel. Outside of the erosional channel, walls and
objects were better preserved. In the rear of the
building only the room [EE], the western roof of
which had already collapsed under the pumice
lapilli (see Section 5.6.1), was invaded from above
by the pyroclastic currents. The remaining rooms,
the roofs of which had remained in place and the
£oor of which had not been covered with the fall
pumice, were invaded through the peristyle (Fig.
14c). No escape was possible for the people who
had taken shelter there. The ash, mixed with
pieces from the roofs, reached every corner of
VOLGEO 2630 21-7-03
G. Luongo et al. / Journal of Volcanology and Geothermal Research 126 (2003) 201^223
221
Fig. 14. The destructive e¡ects of the di¡erent eruptive phases on the Polibius’ house. The section is drawn according to trace I
in Fig. 6; Mount Vesuvius is to the left of the building. (a) Fall products accumulate in the alleys, on the roofs, in the peristyle
and the rooms with impluvium during the early hours of the eruption. (b) Roofs and attics of some front rooms collapse under
the pumice lapilli load. (c) Pyroclastic currents by-pass the rear rooms and destroy the east^west trending walls in the front part
of the house; a re£ected £ow invades the rooms where the inhabitants had taken shelter.
the rooms. A very hard and compact layer formed
under the beds: the skeletons were found lying on
it. In the end, the roofs of the rear rooms also
collapsed, covering the PDCs deposit with a reworked ash and pumice deposit
10. Risk
The present research allows us to draw some
conclusions concerning the impact of explosive
volcanic eruptions onto urban environments.
First, it is possible to state that, as far pumice
fall is concerned, risk is strongly dependent on
the slope of the roofs. Flat and low-angle roofs
are prone to collapse under the pumice load,
whereas steeper roofs tend to drain laterally the
falling material, causing the local accumulation of
thicker deposits (up to 5 m in the case of the AD
79 eruption at Pompeii). The potential danger of
these deposits should be taken into account well
in advance, especially in the case of long lasting
eruptions and/or high accumulation rates. Buildings should be planned so that neither obstruction
of the ways out nor overload of nearby structures
can occur. It is very important to point out that
the roof and wall collapses at Pompeii during the
pumice fall phase of the eruption caused 38% loss
VOLGEO 2630 21-7-03
222
G. Luongo et al. / Journal of Volcanology and Geothermal Research 126 (2003) 201^223
of lives. From a broader point of view this means
that, if roofs are not properly constructed, a relatively high level of damage can be associated
with these kinds of eruptive events. In particular,
as the roofs of the buildings are £at and horizontal in the modern urban areas around Mount Vesuvius, they too will be easily overloaded and collapse. A second consideration applies to the
pyroclastic currents. They had a destructive impact on the structures that happened to strike
perpendicular to their £ow direction, whereas
they little, if at all, a¡ected those that happened
to be parallel. A clear example of this phenomenon in the Polibius’ house was provided by the
surprisingly intact objects found in the closets located along the eastern wall of the peristyle (see
Section 5.3)
Acknowledgements
We are grateful to Professor Guzzo, superintendent at Pompeii for inviting the authors to
participate in the investigation at the Polibius’
house. We also thank Dr. A. Ciarallo and her
collaborators for their logistic help and their interest in our research. We are deeply grateful to
Livio Ruvo for critically reviewing a previous version of this paper. The editorial expertise of B.
Marsh and the helpful comments of two anonymous referees also helped to improve the manuscript. We extend our thanks to Dave Lentz for
his ¢nal revision of the manuscript.
References
Barberi, F., Cioni, R., Santacroce, R., Sbrana, A., Vecci, R.,
1989. Magmatic and phreatomagmatic phases in explosive
eruptions of Vesuvius as deduced by grainsize and component analysis of pyroclastic deposits. J. Volcanol. Geotherm.
Res. 38, 287-307.
Blong, R.J., 1984. Volcanic hazards. Academic Press, New
York, 411 pp.
Branney, M.J., Kokelaar, P., 1992. A reappraisal of ignimbrite
emplacement: Progressive aggradation and particulate to
non-particulate £ow transitions during emplacement of
high-grade ignimbrite. Bull. Volcanol. 54, 504^520.
Bursik, M.I., Woods, A.W., 2000. The e¡ects of topography
on sedimentation from particle-laden turbulent density currents. J. Sediment. Res. 70, 53^63.
Capasso, L., Capasso, L., Caramiello, S., D’anastasio, R., Di
Domenicantonio, L., Di Fabrizio, A., Di Nardo, F., La
Verghetta, M., 2000. Paleobiologia della popolazione di Ercolano (79 d.C.). Recenti Prog. Med. 91, 288^296.
Carey, S., Sigurdsson, H., 1987. Temporal variations in column height and magma discharge rate during the 79 A.D.
eruption of Vesuvius. Geol. Soc. Am. Bull. 99, 303^314.
Cioni, R., Marianelli, P., Sbrana, A., 1990. L’eruzione del 79
d.C.: Stratigra¢a dei depositi ed impatto sugli insediamenti
romani nel settore orientale e meridionale del Somma-Vesuvio. R. St. Pomp. IV, 179^198.
Cioni, R., Marianelli, P., Sbrana, A., 1992. Dynamics of the
A.D. 79 eruption: Stratigraphic, sedimentological and geochemical data on the successions from the Somma-Vesuvius
southern and eastern sectors. Acta Vulcanol. 2, 109^123.
Cioni, R., Sbrana, A., Gurioli, L., 1996. The deposits of AD
79 eruption. CEV^CMVD Workshop on Vesuvius Decade
Volcano, pp. 17^23.
Civetta, L., Galati, R., Santacroce, R., 1991. Magma mixing
and convective compositional layering within the Vesuvius
magma chamber. Bull. Volcanol. 53, 287^300.
Di Girolamo, P., 1964. La serie piroclastica dell’eruzione di
Pompei. Ann. Oss. Vesuv. 5, 1^81.
Druitt, T.H., 1991. Terminology and process in the £ow-surge
debate. CEV Newsletter, 20.
Druitt, T.H., 1992. Emplacement of the 18 May 1980 lateral
blast deposit ENE of Mount St. Helens, Washington. Bull.
Volcanol. 54, 554^572.
Druitt, T.H., 1998. Pyroclastic density currents. In: Gilbert
and Sparks (Eds.), The Physics of explosive volcanic eruptions. Geological Society. Special publication, 145: 145^182.
Fisher, R.V., Heiken, G., 1982. Mt. Pele¤e, Martinique, May 8
and 20, 1902, pyroclastic £ows and surges. J. Volcanol. Geotherm. Res. 13, 339^371.
Fisher, R.V., 1990. Transport and deposition of a pyroclastic
surge across an area of high relief: The 18 May 1980 eruption of Mount St. Helens, Washington. Geol. Soc. Am. Bull.
102, 1038^1054.
Freundt, A., Schmincke, H.-U., 1995. Eruption and emplacement of a basaltic welded ignimbrite during caldera formation on Gran Canaria. Bull. Volcanol. 56, 640^659.
Freundt, A., Bursik, M., 1998. Pyroclastic £ow transport
mechanisms. In: Freundt, Rosi (Eds.), From Magma to
Tephra, pp. 173^246.
Freundt, A., Wilson, C.J.N., Carey, S., 2000. Ignimbrites and
block-and-ash £ow deposits. In: Sigurdsson (Ed.), Encyclopedia of Volcanoes, pp. 581^600.
Kent, D.V., Ninkovich, D., Pescatore, T., Sparks, R.S.J., 1981.
Palaeomagnetic determination of emplacement temperature
of Vesuvius AD 79 pyroclastic deposits. Nature 290, 393^
396.
Lirer, L., Pescatore, T., Booth, B., Walker, G.P.L., 1973. Two
Plinian pumice-fall deposits from Somma-Vesuvius, Italy.
Geol. Soc. Am. Bull. 84, 759^772.
Lirer, L., Munno, R., Petrosino, P., Vinci, A., 1993. Tephro-
VOLGEO 2630 21-7-03
G. Luongo et al. / Journal of Volcanology and Geothermal Research 126 (2003) 201^223
stratigraphy of the A.D. 79 pyroclastic deposits in perivolcanic areas of Mt. Vesuvio (Italy). J. Volcanol. Geotherm.
Res. 58, 133^150.
Lowe, D.R., 1982. Sediment gravity £ows, II. Depositional
model with special reference to the deposits of high-density
turbidity currents. J. Sediment. Petrol. 52, 279^297.
Lowe, D.R., 1988. Suspended-load fallout rate as an independent variable in the analysis of current structures. Sedimentology 35, 765^776.
Luongo, G., Perrotta, A., Scarpati, C., 1995. Vesuvio. L’eruzione del 79 non e' avvenuta all’improvviso. Perche' ? Napoli,
pp. 1^8.
Luongo, G., Perrotta, A., Scarpati, C., De Carolis, E., Patricelli G., Ciarallo, A., 2003. Impact of the AD 79 explosive
eruption on Pompeii, II. Causes of death of the inhabitants
inferred by stratigraphical and areal distribution of the human corpses. J. Volcanol. Geotherm. Res. (in press).
Mastrolorenzo, G., Petrone, P., Pagano, P., Incoronato, A.,
Baxter, P. J., Canzanella, A., Fattore L., 2001. Herculaneum
victims of Vesuvius in AD 79. Nature 769.
Ramsay, P.J., Smith, A.M., Mason, T.R., 1996. Geostrophic
sand ridge, dune ¢elds and associated bedforms from the
Northern KwaZulu^Natal shelf, south-east Africa. Sedimentology 43, 407^420.
Rittmann, A., 1950. L’eruzione vesuviana del 79. Studio mag-
223
matologico e vulcanologico. Pompeiana, raccolta di studi
per il II centenario degli scavi di Pompei, pp. 457^474.
Santacroce, R., 1987. Somma-Vesuvius. C.N.R. Quad. Ric. Sc.
114, 1^251.
Schumacher, R., Schmincke, H.-U., 1991. Internal structure
and occurrence of accretionary lapilli ^ A case study at
Laacher See Volcano. Bull. Volcanol. 53, 612^634.
Sigurdsson, H., Carey, S., Cornell, W., Pescatore, T., 1985.
The eruption of Vesuvius in A.D. 79. Nat. Geogr. Res. 1,
332^387.
Sohn, Y.K., Chough, S.K., 1989. Depositional processes of the
Suwolbong tu¡ ring, Cheju Island (Korea). Sedimentology
37, 1115^1135.
Valentine, G.A., 1987. Strati¢ed £ow in pyroclastic surges.
Bull. Volcanol. 49, 619^630.
Valentine, G.A., 1998. Damage to structures by pyroclastic
£ows and surges, inferred from nuclear weapons e¡ects.
J. Volcanol. Geotherm. Res. 87, 117^140.
Verita', M., Pro¢lo, B., Cancelliere S., 2001. Analisi di reperti
vitrei. In: ‘La casa di Giulio Polibio, Studi Interdisciplinari’,
Ministero per i Beni e le Attivita' Culturali, Centro Studi
Arte Figurative, Universita' di Tokio, pp. 155^162.
Yokoyama, T., Marturano, A. 1997. Volcanic products of the
Vesuvius eruption in AD 79 at Pompeii, Italy. In: Opus.
Pomp., VII, pp. 1^32.
VOLGEO 2630 21-7-03